Regional Differences in Growth Rates: A Microdata Approach

The aim of the present study is to analyse the dynamics of regional consumption and income to explain two significant empirical evidences that have characterized Italian economy in the last two decades: (i) the fall in private saving rate; (ii) the persistence of a wide gap between consumption and income levels of Central-Northern and Southern areas of Italy. The theoretical framework adopted to investigate the effects of economic growth on saving is based on the life cycle hypothesis (LCH) (Modigliani and Brumberg, 1954). As highlighted by recent empirical works, the effect of economic growth on individual saving rates strictly depends on how labour income is affected by growth (Deaton and Paxson, 2000). In this study, we provide a measure of the impact of productivity changes across generations both at the aggregate level and among regions, by tracking income and consumption behaviours of cohorts of households. Moreover, working with household rather than individual data, we adopt an appropriate equivalence scale in order to account for the different resources and needs of each family member; this problem is particularly significant for countries like Italy, in which the presence of multigenerational household is common. The empirical analysis is based on a series of repeated cross-sections of the Bank of Italy’s Survey of Household Income and Wealth (SHIW) for the period 1989-2002 and consists in the decomposition of the cohort, age and time effects of household’s income and consumption along the line of the works of Attanasio (1998), Jappelli and Modigliani (1998), Jappelli (1999) and Kapteyn et al. (2005). The results obtained in the benchmark model show an increase in the productivity of younger generations in the Central and Northern regions together with a positive and increasing age profile for consumption, while in the South the results are floating. The basic model is successively extended by including the demographic and socio-economic characteristics of the household. From the sensitivity estimations, it clearly emerges that household composition, working status and education levels significantly affect income fluctuations in the South, playing an active role in determining the persistency of growth rate differences among regions.